When Should I Do Usability Testing?

In the last post I explained what a usability study is. If you haven’t read that, check it out here.  

Today I wanted to focus on when to do a usability study, but before we dive into that let me first address what types of products benefit from a usability study.

When I talk about usability studies I’m typically focused on tech products which I group into websites, web applications, native applications or mobile apps. I do believe there is benefit in bringing in the user for other types of products during the entire lifecycle and getting feedback along the way as you design and build, whether that be something physical, like the next great widget or something like a book, which Ash Maurya lays out in great detail in his book Running Lean (for that very book).

Bringing it back to tech products my answer would be that they all benefit from usability studies. I like to focus on the low and high end of complexity as use cases and letting the middle sort itself out, so in that spirit let’s break it down:

Low complexity
A portfolio website to show off your artwork and increase consulting work. Surely you don’t need to do any usability testing. It’s straightforward, here’s some beautiful design work and here’s how to get in touch with you. Boom, done. My counter argument to that line of thinking is, what if you could increase the amount of consulting work you get by 10% with a few hours of work? How about 50%? By putting the site in front of people with the task of hiring a design consultant you could discover that they want to fill out a form rather than call or email you. That small change could increase your rate of return.

High complexity
An advertising platform that allows advertisers and publishers to manage their accounts in real time. This would be a web application with a mobile component that allows  both external and internal users different levels of access. With this type of product you would have many levels of usability testing. From a persona level you have advertisers and publishers, all of which are probably not the same and would need to be broken down into sub groups. Even within those personas it’s often the case that users will need different levels of access, such as the Sales Executive and the Ad Buyer. That’s without yet considering how the internal account managers will need to impersonate users and have their own switches and levers.  

Putting these products in front of users is crucial to saving time and money as you’re building and maximizing profits after it hits the market.  

But when do you put it in front of users? When is the right time to do usability testing?

The short answer is, throughout your entire process, from discovery to design, through development and even after launch.

In Tom Chi’s talk on Rapid Iteration (which I posted back in December) he says, “For everyone who builds products, the real medium is human behaviour in the real world, and your purpose should be…to see that human behaviour modified by what you build.”

The longer answer is that the more usability testing you can achieve during your product creation, the better. In Tom Chi’s talk he mentions a feedback loop of only a few days, constantly giving feedback to the design and engineering teams. That is a high bar to reach for indeed.

Instead, breaking the project down into Discovery, Design, Development and Launch we can start to identify the points at which usability tests can be effective, even if we aren’t on a bi-weekly cycle with our feedback loop.

Product discovery starts with understanding what you’re hoping to achieve. See my post on setting your product vision. Involving users at this stage typically means conducting user interviews or nothing at all. One thing you could do as you’re qualifying your intuition is put users in front of similar products and gain insight from how they interact.

During the design phase I encourage iterative design and adding in parallel design if you have the time. Iterative design is the process of taking user feedback on a design and updating it with marked usability improvement with each iteration that you do.  Parallel design starts with multiple concepts and takes the best from all three to end up at 1 design to start the iterative process from.

You want to work through as many cycles as you can, continually gaining insight from users, before a line of code is ever written. In this great article on the Nielsen Norman Group website it goes through the details of each of these projects and state “there’s no one perfect user interface design, and you can’t get good usability by simply shipping your one best idea. You have to try (and test) multiple design ideas.”

During this phase UX, design and engineering should be talking every day about progress of the project and UX should be thinking about what they’re going to test on a weekly basis. In the video from Tom Chi above his team does it twice a week. Do what you can, if you can only do it once every two weeks it’s better than not at all. By taking what engineering is building and putting it in front of users and finding the flaws (and really awesome stuff) as you’re building it will save you from huge rewrites and updates after launch which saves your team huge amounts of time and money.

As I wrote about a few days ago, when you launch you’re not finished, keep pushing towards your goal that you set up at the beginning of the project and while now you will have quantitative data coming in through the many users you have hitting your app or site on a daily basis, it’s also good to sit with your customers and understand how they’re using it in the real world and continue the loop of improving so you can more quickly hit your goals.

In the next post I’ll discuss How to Conduct a Usability Test on the cheap and if you’re interested in having someone take care of it for you, contact me and maybe we can work together.

Leave a Reply